INDEX

Explanations

instances where some information or secret is revealed or discovered

oai_token-act-pair · gpt-3.5-turbo

instances of the word "uncovered" in relation to discoveries or revelations of facts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ente

-0.77

ofi

-0.73

eral

-0.68

erent

-0.68

req

-0.67

jab

-0.67

pop

-0.65

emic

-0.65

enter

-0.65

cas

-0.64

POSITIVE LOGITS

 uncovered

1.07

 unearthed

0.94

çīĪ

0.90

lehem

0.81

 uncover

0.79

ĸļ

0.79

è»

0.77

 mole

0.72

arthed

0.71

 Parenthood

0.70

Activations Density 0.010%

instances where some information or secret is revealed or discovered

instances of the word "uncovered" in relation to discoveries or revelations of facts

No Comments

No Known Activations